Structure of persistently prominent stocks in financial dynamics

Abstract

Prominent components of financial markets have been identified in previous studies using random matrix theory. However, these studies are typically conducted based on static periods. Although these components may dominate during certain periods, they may not necessarily maintain dominance. In financial dynamics, it is important to understand how dominant components persist. In this study, we reveal the structure of prominent stocks by determining the stocks that remain dominant, namely persistently prominent stocks, based on eigenvalue and eigenvector analyses of multiple short time windows. Structural persistence is investigated by a temporal correlation, which suggests better structural sustainability of the persistently prominent stocks for the sector mode and a more stable structure of the whole market for the market mode. Furthermore, persistently prominent stocks are found to have a higher cross-correlation than the whole market for the market mode but show a lower correlation in extreme market states and present a business sector effect for the sector mode.